Re: After Vega 56/64 GPU hang I unable reboot system

["Grodzovsky, Andrey" <Andrey.Grodzovsky@amd.com>]

On 12/17/2018 01:51 PM, Wentland, Harry wrote:
> On 2018-12-15 4:42 a.m., Mikhail Gavrilov wrote:
>> On Sat, 15 Dec 2018 at 00:36, Wentland, Harry <Harry.Wentland@amd.com> wrote:
>>> Looks like there's an error before this happens that might get us into this mess:
>>>
>>> [  229.741741] [drm:amdgpu_job_timedout [amdgpu]] *ERROR* ring gfx timeout, signaled seq=28686, emitted seq=28688
>>> [  229.741806] [drm] GPU recovery disabled.
>>>
>>> Harry
>> Harry, Is this ever will be fixed?
>> That annoying `ring gfx timeout` still follow me on all machines with
>> Vega GPU more than year.
>> Just yesterday I blocked the computer and went to sleep, at the
>> morning I found out that I could not unlock the machine.
>> After connected via ssh I saw again in the kernel log
>> `[drm:amdgpu_job_timedout [amdgpu]] *ERROR* ring gfx timeout, signaled
>> seq=32778472, emitted seq=32778474`
>> It means that this bug may happens even it I doing nothing on my machine.
>>
>> Should we wait for any improvement in localization this bug?
>> Because I suppose message `[drm:amdgpu_job_timedout [amdgpu]] *ERROR*
>> ring gfx timeout, signaled seq=32778472, emitted seq=32778474` not
>> contain any useful info for fixing this bug.
>>
> I don't know much about ring gfx timeouts as my area of expertise revolves around the display side of things, not gfx.
>
> Alex, Christian, any ideas?
>
> Harry

Please install UMR and dump gfx ring content and waves after the hang is 
happening.

UMR at - https://cgit.freedesktop.org/amd/umr/
Waves dump
sudo umr -O verbose,halt_waves -wa
GFX ring dump
sudo umr -O verbose,follow -R gfx[.]

Andrey

>
>> Thanks.
>>
>> --
>> Best Regards,
>> Mike Gavrilov.
>>
> _______________________________________________
> amd-gfx mailing list
> amd-gfx@lists.freedesktop.org
> https://lists.freedesktop.org/mailman/listinfo/amd-gfx